AUDIT OF INVASIVE CERVICAL CANCERSquoting NHSCSP Publication No 28 Tel: 08701 555 455 Fax: 01623 724 524 Email: [email protected] Typeset by Prepress Projects Ltd, Perth () Printed

AUDIT OF INVASIVE CERVICAL CANCERS

NHSCSP Publication No 28December 2006

Published by:

NHS Cancer Screening Programmes Fulwood HouseOld Fulwood RoadSheffieldS10 3TH

Tel: 0114 271 1060 Fax: 0114 271 1089 Email: [email protected] Web site: www.cancerscreening.nhs.uk

© NHS Cancer Screening Programmes 2006

The contents of this document may be copied for use by staff working in the public sector but may not be copied for any other purpose without prior permission from the NHS Cancer Screening Programmes.

ISBN 978 1 84463 040 0

Further copies of this publication are available from the Department of Health Publications Orderline quoting NHSCSP Publication No 28

Tel: 08701 555 455 Fax: 01623 724 524 Email: [email protected]

Typeset by Prepress Projects Ltd, Perth (www.prepress-projects.co.uk) Printed by Duffield Printers

NHSCSP December 2006 iii

Audit of Invasive Cervical Cancers

CONTENTS

Page No

ACKNOWLEDGEMENTS iv

1. INTRODUCTION 1

1.1 Aim of this publication 11.2 Evaluating the effectiveness of the NHSCSP 11.3 Local audit arrangements 2

2. ARRANGEMENTS FOR AUDIT OF CERVICAL CANCERS 3

2.1 Background 32.2 Aims and objectives 32.3 Roles within audit 42.4 Role of cancer registries 82.5 National level 9

3. REVIEW PROCESSES 10

3.1 Screening history review 103.2 Control groups 113.3 Cytology slide review 113.4 Review of gynaecological management 143.5 Histological slide review 143.6 Regional panels 153.7 Reporting of review findings 163.8 Audit of treatment 17

4. ANALYSIS 18

4.1 Evaluation 184.2 Epidemiological audit 18

REFERENCES 19

APPENDIX 1: NATIONAL AUDIT DATASET 20

APPENDIX 2: CERVICAL CANCER CATEGORISATION 37

APPENDIX 3: DIFFICULTIES IN THE IDENTIFICATION OF DYSKARYOSIS 39


NHSCSP December 2006 iv

ACKNOWLEDGEMENTS

The NHS Cervical Screening Programme would like to thank the following for their contribution to the development of this publication:

Members of the Advisory Committee on Cervical Screening

Members of the NHSCSP Laboratory Group (Chair Dr Paul Cross) and the Audit Working Party:

Mr Phil Bullock Advanced Practitioner in Cervical Cytology, Gloucestershire Royal Hospital, Gloucester

Dr Paul Cross Consultant Pathologist, Queen Elizabeth Hospital, GatesheadDr Alistair Deery Consultant Cytologist, St George’s Hospital, LondonDr Keith McCarthy Consultant Histopathologist, Cheltenham General Hospital, CheltenhamDr Bill Young Clinical Scientist Cytology, Hull Infirmary, Hull

Members of the London QA team and the West Midlands QA team

Mr Patrick Walker Consultant Gynaecologist, Royal Free Hospital, LondonDr Terry Rollason Consultant Histopathologist, Birmingham Women’s Hospital, BirminghamSue Hudson IT consultant

Professor Peter Sasieni Cancer Research UKDr Alejandra Castañon Cancer Research UK

Mr Chris Carrigan National Cancer Registry Coordinator for England

The editor is Mrs Julietta Patnick, Director, NHS Cancer Screening Programmes.

NHSCSP December 2006 �


�. INTRODUCTION

The purpose of cervical cancer audit is to monitor the effectiveness of the screening programme and to identify areas of good practice and where improvements can be made. Audits yield information at a national, local and personal level, and the findings consist of the patterns that emerge when the results of the audits of individual cases are analysed together. The aim of this publication is to define a national protocol for audit of cases of invasive cervical cancer in order that standardised data can be pooled and analysed meaningfully.

Judgement about the effectiveness of the NHS Cervical Screening Pro-gramme (NHSCSP) depends on accurate data on incidence and mortality from cancer registries linked to individual level information regarding screening uptake and outcome. Data should be validated and consistently reported, and all parties in the NHSCSP should follow these guidelines. Audit has influenced practice both at bench and clinic levels and in terms of policy development. Sometimes, findings are answers to known ques-tions, and sometimes further questions are identified. There are also times when findings are quite unexpected. This audit will allow future policy and practice to follow the evidence, will allow the promotion of good practice and may identify areas where further attention is needed.

The objective of the NHSCSP is to reduce the incidence of, and mor-tality from, invasive cervical cancer. For women aged 25–64 who are screened in the UK every 3–5 years, it is estimated that cervical screen-ing prevents 75% of invasive cervical cancers by detecting and treating cervical abnormalities which, if left untreated, place women at high risk of developing invasive disease.1

In order to ascertain whether the programme is achieving its objectives, various evaluations are carried out. In particular, the cervical cancer incidence and mortality rates are monitored closely. These show that in recent years the NHSCSP has been very successful. Cervical screen-ing by the NHS in England reduced the incidence of cancer from 15.4 per 100 000 in 1986 to 9.6 per 100 000 in 2000, and increased the rate at which mortality fell from 1–2% per year to 7% per year in 1995.2 Although this rate has since decreased to 5% per year, mortality is now 3.5 per 100 000 (in 2004).3 There are now fewer than 2500 cases of cervical cancer each year, and fewer than 1000 deaths.

However, incidence and mortality alone do not give the complete pic-ture of the programme’s effectiveness. They depict how effective the programme is, not how effective it could be if its activities were all optimised. Audit of the programme will provide this information. Fur-thermore, because cervical screening by means of a Papanicolaou sample has never been subjected to the randomised controlled trials that are today’s gold standard, there are also many questions about its effective-ness that can be answered only by auditing the operational programme in different ways.

�.� Aim of this publication

�.2 Evaluating the effectiveness of the NHSCSP


NHSCSP December 2006 2

Women who develop invasive cervical cancer despite participating in the programme often wish to know why this has happened. Audit of their personal history can yield such information and can provide valuable information on population and operational aspects of the programme. In addition, review of events and specimens from previous years can high-light valuable learning points for the health professionals. The results of such activity nationwide, collected over several years, will yield a great deal of information about the effectiveness of cervical screening.

Although national data collection will enable policy makers to determine whether current policy and practice is working effectively to reduce the morbidity and mortality toll of cervical cancer, local trusts will also be able to use audit to monitor and improve their own practice. It is recom-mended that cervical cancer audit is taken on as part of the organisational clinical governance arrangements in each trust and that each individual/team/directorate within a trust, as well as trust management, determines how audit should work in their own structure. Both reporting the results of audit and learning local lessons from audit should be incorporated into each trust’s own clinical audit framework and clinical governance arrangements.

�.3 Local audit arrangements



2. ARRANGEMENTS FOR AUDIT OF CERVICAL CANCERS

There are often several reasons why women develop invasive cervical carcinoma in a country with an effective population based screening programme. These reasons were recognised before the NHSCSP was implemented in 1988, and were taken into account in previous recom-mendations for audit.4 It is very likely that the majority of cancers detected in the screening age group will occur in women who have previously been screened at some time during their lives because five year coverage has been more than 80% since 1993.

This guidance is not intended to replace local audit and review practice. Frequently, clinicians review women’s management and specimens for their own professional education. Often, if women have been seen at colposcopy, and particularly if they have been treated, this is carried out on a multidisciplinary basis. Where this is the practice, clinicians are encouraged to continue with these reviews, although care should be taken not to compromise the formal audit described here.

Audit of cervical cancer, irrespective of FIGO (International Federation of Gynecology and Obstetrics) stage, is a multidisciplinary procedure involving all elements of the NHSCSP.5 Coordination is vital at all levels to ensure that information is gathered and correlated in a timely and accurate manner. Health professionals who diagnose cervical cancer cases must identify these cases to their hospital based programme coor-dinator (HBPC).

The success of the NHSCSP is reflected in the falling incidence and death rate from cervical cancer. Monitoring the failure of the programme to prevent cervical cancer is also important, increasingly so as changes are made to the technology used and to the age and frequency with which women are called for screening. The national audit of cervical cancers will provide a contemporaneous pattern of disease incidence, including data not recorded by the cancer registries. It will offer the opportunity to explain why some cases occurred, for example in previously unscreened women or if colposcopic treatment has failed, and what proportion were screen detected. It will also, for example, be capable of indicating in a timely fashion whether the alterations in the screening ages and frequen-cies have affected the incidence of cervical cancer. All cervical cancers should be included in the audit, irrespective of clinical stage or the age of the woman at the time of diagnosis.

The aims of national audit are to:

• provide educational feedback to all those involved in the NHSCSP• contribute to monitoring of changes introduced to the NHSCSP• provide a further improvement in cervical screening by identifying

areas of good practice and where the programme may be failing.

2.� Background

2.2 Aims and objectives



The objectives of national audit are to:

• identify screening uptake in women who developed cervical cancer

• have accurate comprehensive data on the disease that essentially represent the outcome of the screening programme

• develop a protocol driven by the quality assurance (QA) offices, based on standard reporting systems

• identify where systematic improvements may be made in national policies, laboratories, call and recall systems and colposcopy clinics

• compare screening histories of those women who have cervical cancer with those women who do not.

The audit will collate regional data on a national basis and will be published annually alongside other nationally collected data for regular analysis. It is proposed to present the findings at an annual meeting. Anonymised data will be stored in a national database. Details of the national audit dataset and arrangements for data collection are given in Appendix 1.

The basic outline and sequence of events in audit are described in Figure 1, and an audit process map is shown in Figure 2. Certain individuals within the programme have been identified as having key roles. Cases of invasive cervical cancer may be identified from a number of sources, including histology laboratories, gynaecology clinics, genitourinary medicine clinics, oncology clinics and cancer registries. When a case is identified clinically (and confirmed histologically), the clinician treating the woman should ensure that the HBPC and the regional quality assur-ance reference centre (QARC) are informed in order for a cascade of audit activities to begin. There are several key roles in the audit process, some of which may be fulfilled by the same individual, eg the HBPC and lead consultant in cytology may be the same individual. Delegation of the roles is acceptable, but responsibility remains with the identified individual and the delegated roles must be identified and agreed by all parties.

The roles can be summarised as follows.

• Report to treating gynaecologist/oncologist when the review process of the patient is initiated.

• Identify the units involved in the patient’s screening history.• Request that local review processes are instigated in all laboratories

holding cytology or histology cases via the appropriate HBPC (generally those cytology laboratories involved within the previous 10 years).

• Request that local review processes are instigated in all units holding colposcopy histories.

• Request the patient’s screening history from the Exeter database managers as appropriate.

• Report the outcome of the completed review to the treating gynaecologist/oncologist.

2.3 Roles within audit

2.3.1 Hospitalbasedprogrammecoordinator(HBPC)intrustreportinginitialhistologicaldiagnosis

Figure � Outline of audit sequence.

Screening history reviewed

Cervical cancer identified

Any cytology slide (within previous 10

years) reviewed

Notification to cancer registry of cancer classification

Notification to QARC of cancer classification

Local review of findings and notification to woman if required

Any colposcopy notes reviewed

Review of GP records (if possible) for non-attenders

and uninvited women

National cancer audit entry (with controls)

Any histology specimens reviewed

Figure 2 NHSCSP prospective audit of cervical cancers audit process map.

Cancer identified to HBPC in trust where cancer is diagnosed

HBPC notifies treating gynaecologist or oncologist

that review process has been instigated and offers to report

findings

Treating gynaecologist or oncologist reports review

findings to patient

Registered GP extracts history

from notes

Exeter database managers

extract screening

history

Lead consultant in cytology

instigates local review process

Lead consultant in histology

instigates local review process

Lead consultant in colposcopy instigates local review process

Local review outcomes do not confirm original

report

Review findings equivocal

Local review outcomes

confirm original report

Review findings unequivocal

Local review findings reported

back to HBPC in trust where

cancer is diagnosed

HBPC in trust where cancer is diagnosed

reports findings to treating

gynaecologist or oncologist

HBPC in trust where cancer is diagnosed reports reviews to QARC quarterly and

local screening programme commissioner annually

Local review findings and

materials sent to QARC for panel

assessment

HBPC identifies trusts, Exeter databases and registered GP holding

case history and requests reviews

HBPC notifies QARC in own

region that review process

has been instigated

QARC panel assessment

findings reported back to HBPC in trust where

cancer is diagnosed



• Assist with feedback to the patient as required.• Ensure a fail safe system is in place in local laboratories to identify

all cervical cancers.• Report completed reviews using standard forms to the QARC (see

Appendix 1).• Notify the local screening commissioner of audit outcomes on an

annual basis.

• Instigate the local cytology review process.• Report the outcome of review to the coordinating HBPC using the

standard forms.• Refer appropriate cases to QARC for further review via the trust

HBPC.

• Instigate the local histology review process.• Report the outcome of review to the coordinating HBPC using the

standard forms.• Refer appropriate cases to QARC for panel review via the trust

HBPC.

• Instigate the local colposcopy review process.• Report the outcome of review to the coordinating HBPC using the

standard forms.• Refer appropriate cases to QARC for panel review via the trust

HBPC.

• Supply screening histories to HBPC using the standard forms.• Identify controls.• Supply information on controls, including screening histories, to

HBPC using the standard forms.

• Carry out a local audit of own practice (practices that participate in the programme may be remunerated through the Quality and Outcomes Framework (QOF)).

• Cooperate with wider audit protocols as necessary for patients in the practice.

• Validate local cytology, histology and colposcopy review processes in line with this protocol.

• Convene cytology, histology and colposcopy review panels to review difficult cases.

• Report the outcome of case reviews back to HBPC in all trusts involved in order to feed back to lead consultants in the trust for cytology, histology or colposcopy.

• Monitor progress and outcome of audits.

The QARC input is regional coordination of the cervical cancer history review. This will involve all elements of that review, and will also ulti-mately produce the national database return and assist in the collection of data regionally.

2.3.2 Leadconsultantincytologyinalllaboratoriesholdingcytologyslides

2.3.3 Leadconsultantinhistologyinalllaboratoriesholdinghistologycases

2.3.4 Leadcolposcopistinallunitsholdingcolposcopyhistories

2.3.5 Callandrecallsystemmanager

2.3.6 Primarycareteam

2.3.7 QualityassuranceteamandQARC



The data collected via the HBPC and passed on to the QARC will be submitted nationally by the QARC using the standard data format (see Appendix 1).

• Receive audit data annually from HBPC and incorporate into annual reports.

• Work with relevant primary care staff (GPs, practice nurses, community clinics, etc) to contribute to the audit review process as outlined above.

Cancer registries are responsible for the identification and registration of cervical intraepithelial neoplasia 3 (CIN 3) and invasive cervical cancers diagnosed in women resident in their catchment area, and for continued collection and analysis of diagnosis and treatment data that follow diag-nosis. Cancer registries are required to categorise the diagnostic route for each invasive cervical cancer, including screening status, for the National Cancer Dataset.

All cervical cancers should be categorised by the QARC into one of six categories, which in turn map to the diagnostic status required by the National Cancer Dataset used by the cancer registries. Details of the categories of cervical cancer and the requirements of the National Cancer Dataset are given in Appendix 2.

The director of each regional cancer registry should identify an individual within that registry as having lead responsibility for liaising with the QARC to ensure that the information stored by the registry includes a record of the diagnostic status (screening/interval/other/not known) of each woman with invasive cervical cancer diagnosed in the population eligible for screening.

This individual should work closely with his/her opposite number at the QARC in order to ensure complete ascertainment and analysis of all cases in women resident in the region covered by the registry and to ensure full exchange of data about women diagnosed or screened in the area but resident within another registry’s area. The role of this lead person includes:

• liaison with the relevant QARC in order to achieve full ascertainment of all CIN 3 and invasive cervical cancers diagnosed in women resident in the region

• supply of data to the QARC for all diagnosed women within the screening population (whether the registry has the screening status or not)

• receiving the results of checking a woman’s NHSCSP history, diagnosis and screening classification from the QARC

• assisting the QARC as required in the analysis of patterns of cases in order to identify any weaknesses that require attention.

2.3.8 Screeningprogrammecommissioners

2.� Role of cancer registries



The national office of the NHSCSP will work closely with the QARC, the cancer registries and Cancer Research UK to collate and analyse data from each individual woman in order to produce information about the sensitivity and performance of the NHSCSP at a national level. This will result in scientific publications and annual data that will illustrate the numbers of women falling into each category. Details are given in Chapter 4. These national data will allow evaluations of changes in policy, year on year comparisons of the performance of the NHSCSP, international comparison of performance and comparison of perform-ance against expectation and trials data in order to improve estimates of performance. In addition to the collation of data from regional QARCs, working closely with the cancer registries will allow classification of the cancers detected into various diagnostic categories to aid evaluation of the programme. The epidemiological audit, in collaboration with Cancer Research UK, will evaluate the programme’s policies and how effectively they are implemented. The epidemiological audit will require audit of a control group.

2.5 National level


NHSCSP December 2006 �0

3. REVIEW PROCESSES

The full screening history for a particular woman can be obtained only by searching a number of different databases. The first step should be to use the local computerised call/recall NHAIS computer system (the Exeter system), which holds a woman’s invitation and cytology history. Only when the history is ascertained can the next steps be identified.

The history will require demographic details (name, date of birth, etc) to allow a search on the call/recall, laboratory and colposcopy computer systems. It should be possible to do this through the local call/recall office because all screening invitations, results, final non-responder cards, etc would be detailed. Access may also be required to the ‘deducted’ (removed) patient details.

The assistance of the local screening manager (or equivalent) and the local screening commissioner/primary care trust (PCT) and HBPC may also be required to allow the full screening history to be obtained. If the HBPC has access to the Open Exeter web browser or the new Patient Demographics Service through ‘Choose and book’, this may allow an initial search. However, as access to the Exeter database varies locally, each area should determine who is the most appropriate person to under-take the actual search to ensure that all records are identified.

The screening history obtained from the search outlined above should identify the full history of screening invitations, results, invitations not acted upon, etc as well as the laboratories involved in reporting the cytology and the source of sample. The latter would facilitate finding colposcopy or GP records. This Exeter derived screening history will form the basis of the screening history for the purposes of the cervical cancer audit.

All local colposcopy clinics should be contacted for relevant records. The record should include the dates of all appointments, whether the patient attended and whether any procedures were carried out (see Appendix 1, section C). It should also include colposcopic impression and treatment.

A record of histology results should be collated to produce a complete picture of the patient’s history and to facilitate slide review. Information should be collected as detailed in Appendix 1, section D.

It may also be desirable to check GPs’ notes on certain patients, particu-larly to understand any reasons for non-attendance. Such information, if obtained, should be recorded and should form part of the patient’s audit.

Other episodes may be identified within non-NHS facilities, and access to records for this may also have to be obtained by the HBPC or QARC depending on local circumstances.

3.� Screening history review

NHSCSP December 2006 ��


In order to allow rigorous evaluation of the programme, women who did not develop cancer will be required as controls. This exercise will be carried out in cooperation with Cancer Research UK. For each woman with invasive cancer, two women who have not undergone hysterectomy should be matched for age and area of residence. The selection of controls can now be carried out automatically using the NHAIS (Exeter) system. This facility will be incorporated into the new national system that will replace NHAIS under the National Programme for IT being developed by NHS Connecting for Health. Cancer Research UK will issue separate guidance on selection of controls for the epidemiological audit.

Cytology slide review is a powerful tool for both audit and education. For this reason, it is best performed within the laboratory that reported the original slides. In the event of tests having been reported in more than one laboratory, each laboratory should review its own slides and forward the results to the HBPC at the laboratory generating the original histological diagnosis of cervical cancer. In the event of this not being a cervical screening laboratory, this responsibility should pass to the HBPC at the trust laboratory reporting the most recent test.

Any audit must be constructed so as to maximise the educational value for the NHSCSP as a whole. Cytology slide review should be carried out only with current staff with current knowledge of cervical cytology reporting and of its pitfalls. Although it is accepted that reporting and diagnostic criteria may have changed in the interval since the slides for review were reported, any attempt to replicate historical working prac-tices is fraught with problems and should not be made. Difficulties in the identification of dyskaryosis are discussed in more detail in Appendix 3, which has been adapted from previously published advice.4

The aim of a slide review audit is not to replicate ‘normal’ screening practice, but rather to identify lessons for the NHSCSP as a whole. In light of this, the slide review does not have to follow traditional screen-ing pathways in that it is not a ‘test’ of whether a slide should have been reported differently by different grades of staff. The purpose is to see whether there were reasons why the particular cancer in question developed, and whether there were any cytological reasons that may have contributed to this. The review is not a medicolegal review and is carried out by NHS staff for educational purposes.

The histological diagnosis of a cervical cancer is the event that will trig-ger a review of that woman’s cervical screening history. The pathological aspect of this will require a full review of the cytology and histology history. Histology review is described in section 3.5.

The sample history can be obtained from the Exeter database either via Open Exeter or from the call/recall office. This will allow identification of where and when slides were reported. If this involves other laboratories, the HBPC at the hospital making the histological diagnosis will write to the HBPC at any other hospitals where cytology has been reported

3.2 Control groups

3.3 Cytology slide review

3.3.1 Generalapproach



to ask for a full screening history review to be undertaken in line with this protocol, including slide review. If these hospitals are no longer screening laboratories within the NHS or the laboratory is outside the NHS, the initiating hospital will offer to undertake this review. If the hospital making the original diagnosis is not a screening laboratory, the hospital with the most recent sample that is an NHSCSP laboratory will undertake that review.

The slide review process must have access to the original report as issued and to the original request/report card wherever possible. Any laboratory records or ‘in house’ comments should also be available for review.

Many slides will have screening dots on them already. It is possible that as part of the review new ones may be added to identify areas of interest. To aid in this, it is recommended that a copy of the slide is taken prior to any review (often a simple photocopy or use of a slide graticule will suffice). If new dots are added, then the slide should be copied again after the review. The dots will aid in any discussion/education based on slide review as part of the audit process.

The internal laboratory review process must entail a review of the original slides. A model for this is outlined below. Figure 3 outlines the process as a flow diagram.

• The slides identified will be reviewed independently by two members of the screening staff, one a screener and one a checker, and the results recorded on the data collection sheets shown in Appendix 1. Given that slides should be kept for at least 10 years, this should not produce a large workload for any one laboratory.

• Once completed, the slides and forms are given to a consultant pathologist or advanced biomedical scientist practitioner currently reporting samples within the NHSCSP for a third review and for collation of the three slide reviews. This will identify slides where there is agreement or non-agreement, either with the original report as issued or with the review opinions themselves. Discussion should also take place between the reviewers to aid in this process, and to maximise the educational value for the reviewers involved.

• The information is then passed to the regional QARC for entry into the regional, and then national, cervical cancer audit database via the HBPC in the laboratory initiating the review. Any cases where there is a lack of agreement following case discussion, where all review slides are classed as non-dyskaryotic or where the reviewing laboratory asks for a further opinion are referred to the QARC for panel and/or expert review as considered appropriate. The process for review should, in most cases, be completed within 6–8 weeks of the original histological diagnosis.

• If any cytology slides are not available for review, this must be noted on the QARC return.

• The reviews undertaken by other laboratories apart from the initial hospital must be returned to the HBPC at the initiating hospital within one month of receipt, and this information is then passed to the regional QARC of the initiating hospital.

3.3.2 Internallaboratoryreview

Is the slide technically suitable for review (stain not faded, not

dried back)?

Slide should be rejected as unsuitable for review

Assess cytological appearances

Yes

Does the appearance of the sample fall within

normal limits?

Comment on degree and type of abnormality present

Comment on borderlines

Yes

Negative result

Bearing in mind knowledge at the time of the original report, does the slide fall into one of the known ‘difficult to

identify’ categories?

Estimate number of abnormal cells present

Approximately > 200

Approximately < 200

Scanty dyskaryosis: possible/probable

unavoidable false negative

Small cell dyskaryosis, pale cell dyskaryosis, microbiopsies, other

Yes

Potential difficult diagnostic category

No

Possible/probable unavoidable false negative report

Potentially avoidable false negative report

No

No

Figure 3 Slide review decision making.



• All reporting must use accepted NHSCSP/British Society for Clinical Cytology terminology.4

• The cytological review will record data not only on the review opinion but also on type, number and appearances of dyskaryosis (if present).

Where review of gynaecological/colposcopic management is required, this should be undertaken by two consultant gynaecologists who are accredited colposcopists. They should review any colposcopic assess-ment for which records are available.

Unlike cervical cytology and histopathology, there has traditionally been no permanent record kept of a colposcopic image taken at the time of assessment. It has always been good practice to record three features in patients’ records:

• whether the new squamocolumnar junction is visible• whether there is an abnormality present• an impression of the degree of abnormality.

These three features are included in the standard for documentation to be recorded in ColposcopyandProgrammeManagement (NHSCSP Publi-cation No 20).6 Most colposcopists produce a small diagram within the record demonstrating these features and recording a management plan. In recent times, the capacity to record a digital image of the colposcopic findings has become standard practice in some clinics and may be useful if available, although the degree of agreement between observers of images can be highly variable. The published guidelines indicate areas of current good practice for colposcopy.

The histology slide review should include slides from which the diagnosis of cancer was made and also any slides from the previous 10 years. Only existing slides should be reviewed – there is no need to cut new sections. The review should also include a macroscopic examination of any blocks if there is any suggestion that all pieces have not been cut into or if there is a clear discrepancy found in the review. Pathologists may wish to cut additional slides to seek further information that might change the grade of CIN originally reported or to identify possible invasion.

Audit of cervical cancer will also include a full review of any relevant histological material. This will include cervical biopsies and cervical excisions (e.g. large loop excision of the transformation zone, loop or knife cones, etc). The identification of such material may be more problematic than that of the cervical sample slides in that there is no one database for this material that is similar to the cytology history held on the Exeter system. However, referral to this database will usually give a good indication of when a woman was referred for colposcopy and therefore when histology specimens were likely to have been taken. The

3.� Review of gynaecological management

3.5 Histological slide review

3.5.1 Generalapproach



primary care record could also be used to identify relevant colposcopy episodes and hence potential histological material.

Histology can be identified by review of pathology laboratory compu-ter or other record keeping systems at all hospitals where slides were reported. Hospitals will know their own referring patterns locally, and this may mean contacting other centres not involved in reporting cervi-cal samples. The colposcopy history review may also identify centres where colposcopy was undertaken, and hence where histological mate-rial may be kept.

The review should use the original slides and have access to the original report as it was issued.

• The histological review can be undertaken only by consultant pathologists. Ideally, these would be consultants who participate in the national gynaecological pathology EQA scheme and who routinely report on NHSCSP histological material. The opinion of one, but preferably two, consultants should be obtained and recorded using the data collection sheets shown in Appendix 1.

• It is important to indicate whether a specimen was received in one piece or piecemeal. The presence or absence of CIN, cervical glandular intraepithelial neoplasia (CGIN), invasive disease and any other relevant pathology, as determined by the review panel, should be recorded, together with an indication of any particular difficulties in the interpretation of the case, eg severe diathermy artefact, epithelial stripping or poor orientation of the sample.

• The review should use standard NHSCSP/FIGO/RCPath terminology.5,7–9

• Once collected, the data must be sent to the QARC of the region of the initiating hospital. This will be via the HBPC of the hospital initiating the review. Demographic data will be collected as detailed on the cytology review form, and histology data as detailed on the RCPath minimum data format form.7

• Cases for which there is a lack of consensus between the reviewing pathologists, or which are perceived as difficult, are referred to the QARC for regional panel review.

The regional QARC will need access to an expert(s) or to a panel for cytology, histology and/or colposcopy review in the following circum-stances:

• cases identified as problematic by the reviewers• cases in which all samples are identified as negative or inadequate

or borderline after local review.

All results should be notified to the HBPC in the initiating hospital by the QARC.

3.5.2 Conductofthereview

3.6 Regional panels



For cytology review, the panels should ideally consist of a minimum of three people. These would include a primary screener and/or a checker and a consultant pathologist and/or an advanced practitioner. They should independently review the cases submitted, and then review their own opinions against the original hospital based review. If this regional review identifies a case that is problematic, this needs to be noted and included in the report to the HBPC in the initiating hospital.

For histology review, the panel should consist of two consultant his-topathologists who routinely report NHSCSP related material, at least one of whom regularly participates in the national gynaecological his-topathology external quality assessment scheme. They should independ-ently review the slides and then compare their opinions with the original hospital based review. If this regional review cannot reach consensus, this needs to be noted and included in the report to the HBPC in the initiating hospital.

The regional review process should be completed within one month of receipt of all the required material.

The regional panel opinion is recorded as the ‘final’ opinion in these reviewed cases. The regional panel view is conveyed to the HBPC in the initiating hospital, and all the other HBPCs involved, by the QARC. A final outcome of ‘equivocal’ is acceptable in cases where agreement cannot be reached even after regional review. The HBPC will communi-cate the review outcome to all other hospitals involved in the review.

Any hospital where an HBPC cannot be identified needs to consult the regional QARC to ensure clarity of communication.

Identifying details, such as the patient’s name and NHS or hospital number, should be covered over so far as is possible without obscuring original detail in order to protect patient confidentiality before QARC review. Such details should not be permanently erased from the slides.

The QARC will collate the information relating to the cervical screen-ing history and complete the national data return for national analysis. This will include not only the slide review but also the colposcopy and screening history review.

The HBPC at all the sites involved in each case will receive a full screen-ing history for each woman relevant to their site from the HBPC initiating the audit process when it has been completed. This can then be passed on to the individual clinicians involved (lead cytologist, lead histologist, lead colposcopist, PCT lead, call/recall lead) for information and could be used as a basis for discussion with the woman if she so wishes.

The report issued constitutes an NHSCSP audit review, not a legal review, and as such must be carefully discussed in this light.

The nationally collected data will allow annual data publication, and can be issued alongside other nationally collected data for regular analysis.

3.� Reporting of review findings



This may allow lessons about the NHSCSP at all points in the screening pathway to be identified, and so may lead to benefits that will help to improve the NHSCSP.

The purpose of this audit is to help to improve the NHSCSP. From an educational stance, this can be maximised within each department involved in the management of women in the programme by:

• discussion of the review results with all screening staff and any other health professional involved

• production of an annual report of the individual departmental experience, findings and any action points resulting (eg training, etc)

• production of an annual QARC/national report of the findings of the review, lessons learned and action points for the programme

• discussion at a local level (laboratory, PCT, colposcopists, etc) of findings, trends, etc.

All new cases of cervical cancer should be audited. All cases logged with the regional QARCs should be cross-checked against those registered with the cancer registries (see section 2.4).

An audit of the treatment of cervical cancer is being planned and guid-ance will be added following piloting. This will use essentially the same dataset as all other audits in the NHSCSP. In time, as information systems improve, it is envisaged that audit of cases of CIN 3 will be possible.

3.8 Audit of treatment



�. ANALYSIS

Women with cancer will be categorised into the following groups:

1. Screen detected cancer 2. Interval cancer 3. Lapsed attender 4. Never invited (subgroups by < 25, 25–64, 65+) 5. Never attended 6. Lost to follow up

Study of the different groups of women by the local QA teams will yield valuable information about where resources to improve the programme or correct deficiencies might best be directed. For example, those women who have never been invited but are however within the screening age group will need particular scrutiny by their local QA team to ascertain the reasons for their non-invitation, and study of those women who have never attended may assist in learning where activities to improve access to the programme might best be directed.

Further detail on these categories is given in Appendix 2.

This audit will use the categories described above but will further sub-categorise those with a cytological history to consider:

1. Screen detecteda. Cancer diagnosed at referral for colposcopyb. Cancer diagnosed after negative cytological follow up at col-

poscopyc. Persistent abnormality after treatment diagnosed at follow up

cytology2. Interval cancers

a. Previously screened as recommendedb. Not previously screened at recommended frequency

3. Lapsed attenders with previous negative cytologya. Previously screened as recommendedb. Not previously screened at recommended frequency

4. Lost to follow upa. Abnormal cytologyb. Referral indicated and not attendedc. Follow up not attended after treatment for CIN

This will enable an evaluation to be made of the policies for follow up and for the age and frequency of invitation. It should be recognised, however, that the proportion of cases in each group reflects the screen-ing coverage of the target population as well as the effectiveness of the screening programme. It is for this reason that the inclusion of controls is essential in the evaluation part of audit activities.

�.� Evaluation

�.2 Epidemiological audit



REFERENCES

1. IARCHandbookofCancerPrevention. No 10. CervicalCancerScreening. IARC, 2005.

2. Sasieni P, Cuzick J, Farmery E. Accelerated decline in cervical cancer mortality in England and Wales. TheLancet, 1995, 346: 1566 [Letter].

3. CancerStatistics. Cancer Research UK, 2006 (www.cancerresearchuk.org). 4. AchievableStandards,BenchmarksforReporting,andCriteriaforEvaluating

CervicalCytopathology. NHS Cancer Screening Programmes, 2000 (NHSCSP Publication No 1, second edition).

5. FIGOStagingforCervicalCancers (www.figo.org/content/PDF/staging-booklet.pdf).

6. Luesley D and Leeson S (eds). ColposcopyandProgrammeManagement. NHS Cancer Screening Programmes, 2004 (NHSCSP Publication No 20).

7. MinimumDatasetfortheHistopathologicalReportingofCervicalNeoplasia. RCPath, 2001 (www.rcpath.org/resources/pdf/datasetrecervicalcancer.pdf).

8. HistopathologyReportinginCervicalScreening. NHS Cancer Screening Programmes, 1999 (NHSCSP Publication No 10).

9. BenchAidfortheReportingofCervicalHistologicalSamplesAssociatedwithAbnormalCervicalCytologyorColposcopicExaminations. NHS Cancer

Screening Programmes, 2003 (NHSCSP Publication No 16).



APPENDIX �: NATIONAL AUDIT DATASET

A�.� Common national database

The use of a common national database will facilitate the pooling of data from screening programmes across the country to allow epidemiological analysis. This is not a minimum national dataset (ie not all fields are essential); please refer to the coding guide for a list of essential fields.

A�.2 Access database

An electronic (Access) database is available on request and is provided with an explanatory manual. Please email [email protected] for a copy or for further details. The intention is that the electronic database will initially be used by QARCs to input the data collected by HBPCs.

A�.3 Data collection forms

HBPCs are expected to use the forms in this Appendix for data collection. Different forms are to be completed by different laboratories or clinics either by a specialist from the relevant area or by the person responsible for the collection of audit data. It is not necessary to complete all sections in the audit in order to submit data; however, sections A and B are essential.

A�.� Coding guide

The coding guide for the national dataset provides an overview and explanation of the sections and fields that the audit aims to record. The last section lists the fields that are essential for audit purposes. The fields that are not mentioned are desirable, ie an effort should be made to collect the data but forms should be submitted even if there are some fields missing.

A�.5 Printable forms and coding guide

The data collection forms and coding guide listed below are included on the following pages. Printable copies of the forms as separate PDFs can be downloaded from the NHS Cancer Screening Programmes website (www.cancerscreening.nhs.uk). A copy of the coding guide can also be downloaded.

Datacollectionforms

Section A Personal and cancer detailsSection B Cytology historySection C Colposcopy historySection D Histology historySection E Cytology reviewSection F Histology reviewSection G GP notesSection H HPV DNA testing

Codingguide

• Personal and cancer details• Cytology history• Colposcopy history• Histology history and review• Cytology review• GP notes• HPV DNA• Essential fields

NHSCSP December 2006 3�


APPENDIX 2: CERVICAL CANCER CATEGORISATION

A2.� Categories of cervical cancer1. Screen detected Detected after a diagnostic process that began with a cytology test taken up to three

months before the test due date or up to six months after the test was due

2. Interval cancer Detected in the interval between test due dates with the previous episode having been closed with no diagnosis of cancer

3. Lapsed attender Cancer detected in a woman who had previously been screened at least once, whose last result was negative and who had not attended when last invited (up to six months after the test was due) and who was less than 70 years old (ie less than five years above the upper age limit for invitation)

4. Never invited Woman who has never been tested nor sent an invitation for screening (subgroups by < 25, 25–64, 65+)

5. Never attended Woman who has never been tested but who has been invited for screening

6. Lost to follow up Woman in whom either colposcopy or repeat smear was indicated, but who never received any follow up

Notes

1. Categories 2–6 ignore any cytology taken as part of the diagnostic process.

2. The screening invitation must be sent by the NHAIS system or its dispatch must be verified by checking with the woman’s GP whether the practice sends its own invitations.

3. Cancers in women who are under follow up after colposcopy and treatment should be categorised according to their test status, eg was their cancer detected when they responded to an invitation to follow up (category 1) or did they fail to attend when invited (category 6)?

4. The test due date will be affected by the woman’s screening status immediately before being diagnosed with cervical cancer (routine, early recall/suspended, post-treatment).

A2.2 National Cancer Dataset, Version �.0 Issue date ��.08.03

C1 Diagnostic route (screening status)

Indicates the patient’s route to diagnosis

1. Cancers detected by the national screening programme (category 1 above)2. Interval cancers occurring in patients screened by a national screening programme (category 2 above)3. Other cancers (categories 3–6 above)9. Not known (default)

All

cerv

ical

can

cers

Nev

er

invi

ted

Scr

een

dete

cted

ca

ncer

s

Rou

tine

call/

reca

ll (in

clud

ing

early

re

peat

s)

Follo

w u

p

Laps

ed

atte

nder

s

< 2

525

–64

65+

Inte

rval

ca

ncer

s

Rev

iew

of s

cree

ning

his

tory

, im

ages

an

d sa

mpl

es a

s ap

prop

riate

Nev

er

atte

nded

Lost

to fo

llow

up

Ref

erra

l for

co

lpos

copy

in

dica

ted

Rep

eat

test

in

dica

ted

Fig

ure

A2.

� C

ateg

oris

atio

n of

cer

vica

l can

cers

.

NHSCSP December 2006 3�


APPENDIX 3: DIFFICULTIES IN THE IDENTIFICATION OF DYSKARYOSIS

It is relatively easy to recognise the classic form of severe dyskaryosis in which the dissociated cells have high nuclear–cytoplasmic ratios and hyperchromatic nuclei with irregularly dispersed chromatin. There are, however, several other cytological patterns indicating the presence of CIN 2/3 that are less easy to recognise and may lead to false negative cytology reports. The following sections draw attention to the main patterns which screeners, biomedical scientists and pathologists should recognise as potential problems.

Small severely dyskaryotic cells may be only the same size as a neutrophil polymorph or even smaller. They sometimes have regular nuclear mem-branes, in which case their correct recognition depends on the apprecia-tion of abnormal, irregularly clumped or speckled chromatin patterns. Nucleoli are usually, but not invariably, inconspicuous. Such cells may be mistaken for histiocytes, lymphocytes, endometrial cells or immature metaplastic cells. The key to recognising these cells is the characteristic nuclear chromatin pattern in association with a high nuclear–cytoplasmic ratio, despite the small size of the cells. Careful searching may reveal cells with keratinisation, confirming their squamous cell type. Many samples with small cell severe dyskaryosis will also include dyskaryotic cells of lesser grade which are obviously squamous in type. The observation of a continuum of cytological features from unequivocal mild or moderate squamous dyskaryosis into a small cell population may help the confident identification of small cell severe dyskaryosis.

Dyskaryotic nuclei are not necessarily hyperchromatic, and dyskaryosis may be seen in deceptively hypochromatic nuclei from all grades of CIN and even invasive squamous cell carcinoma. Pale dyskaryosis is often seen in samples mixed with cells showing more classic or hyperchromatic dyskaryosis; however, when this occurs as the predominant or only type in a sample, its recognition may be particularly difficult. Careful attention to the chromatin pattern, as described above, should allow recognition of this subtype.

Severe dyskaryosis may be seen in sheets or three dimensional aggregates of cells which frequently appear crowded and hyperchromatic, and such aggregates are recognised as a common cause of errors of interpretation (as opposed to detection). They may easily be mistaken for endocervi-cal cells. Diagnostic clues to the presence of severe dyskaryosis include disorderly cell arrangements with loss of polarity or chaotic architecture, mitotic figures (especially if numerous or abnormal) and a coarse, dark chromatin pattern. The last may be particularly difficult to evaluate in three dimensional clusters, and careful attention to the nuclear chromatin and nuclear–cytoplasmic ratio of cells at the edge of the group, especially if single non-overlapped nuclei can be seen, should help in interpretation. Aggregates of small severely dyskaryotic cells, especially if also showing pale dyskaryosis, may be very difficult to interpret. They may appear to

A3.� Potential false negative results

A3.1.1Smallcellseveredyskaryosis

A3.1.2Paledyskaryosis

A3.1.3CIN3‘microbiopsies’andCIN2and3infiltrating crypts



be deceptively orderly; columnar cells may be seen in small cell CIN 3 lesions and the aggregates may even on occasions have a border of low columnar cells. It is very unusual for CIN lesions to present in cervical samples as only cell aggregates without any single dyskaryotic cells, and the observation of dyskaryosis elsewhere in the sample may assist in interpretation. If a confident conclusion cannot be reached, it may be necessary to use the borderline category for reporting, and this is one situation where it may be justifiable for this report to warrant immediate referral for colposcopy.

Severe, and less frequently moderate, dyskaryosis may be intimately associated with endocervical cells in such a way that the cell group may be considered to be entirely glandular. This feature is sometimes taken to indicate crypt infiltration by CIN 2/3. Groups of this type with a columnar edge of apparently normal endocervical cells in places are occasionally seen and present a particular diagnostic pitfall. High power examination of individual nuclei should reveal the characteristics of dyskaryosis in some of the cells. The characteristic architectural features of glandular neoplasia are not seen in these groups.

Small keratinised dyskaryotic cells may be difficult to recognise in atrophic samples, particularly in association with inflammation. If in doubt, an early repeat sample after topical estrogen treatment may be justified, as dyskaryotic cells are often much easier to recognise in a more mature pattern sample.

Sparse severely dyskaryotic cells may be difficult to grade and may be misinterpreted as mild dyskaryosis or borderline nuclear change. The degree of dyskaryosis shown by abnormal cells should not be downgraded because of their scarcity in a sample.

False positive results for samples reported as severe dyskaryosis are unusual. They are more common in samples reported as moderate dys-karyosis or glandular neoplasia in which the abnormal nuclear changes may be less obvious. The following conditions and cell changes occa-sionally give rise to false positive results.

Normal shed endometrial cells may be mistaken for small dyskaryotic squamous cells. Careful attention to clinical data, date of last menstrual period in relation to the sample, intrauterine contraceptive device use or sex hormone therapy and to the sample appearances and cell detail will usually enable correct identification of such cells to be made.

Endometriosis and tubal or tuboendometrioid metaplasia may occur spon-taneously in the cervix, and are likely to occur much more frequently after cone biopsy and other operative procedures. Endometrial stromal cells may mimic dyskaryotic squamous cells, and large combined glandular and stromal or glandular cell groups are more likely to be mistaken for abnormal endocervical cell groups. It should be noted that the nuclei of endometrial and tubal epithelial cells may normally appear to be pseu-dostratified.

A3.1.4Smallkeratinisedcells

A3.1.5Sparsedyskaryoticcells

A3.2 Potential false positive results

A3.2.1Normalendometrialcells

A3.2.2Endometriosisandtuboendometrioidmetaplasia



Endometrial material may be sampled directly, possibly because of shortening of the endocervical canal after treatment but more frequently when endocervical brushes, or other sampling devices for improved endocervical sampling, are used. Such material may include glandular and stromal cells, and often includes ‘microbiopsies’ of endometrial tissue. The recognition of such large biphasic groups is important in the identi-fication of LUS endometrium; if both glandular and stromal cells can be identified in the same cell group, the endometrium is extremely unlikely to be neoplastic. LUS endometrium may respond to hormones, and mitotic activity may be seen in the stromal or epithelial components.

Histiocytes are normally easily recognisable but, especially when they become degenerate, they may show granular or dense chromatin and dense cytoplasm, closely mimicking the appearance of severe squamous dyskaryosis of small cell type. Occasionally, usually in late menstrual samples, the cytoplasm of histiocytes may become eosinophilic, resem-bling keratinisation.

This condition may occasionally be misinterpreted as severe dyskaryo-sis or as endometrial cells. Attention to the typically coarse but evenly clumped chromatin and the presence of tingible body macrophages should determine the correct diagnosis.

A3.2.3Loweruterinesegment(LUS)endometrium

A3.2.4Histiocytes

A3.2.5Follicularlymphocyticcervicitis







Documents

AUDIT OF INVASIVE CERVICAL CANCERSquoting NHSCSP Publication No 28 Tel: 08701 555 455 Fax: 01623 724 524 Email: [email protected] Typeset by Prepress Projects Ltd, Perth () Printed